This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: Make check failure with gcc-4.1 on PPC
- From: Jakub Jelinek <jakub at redhat dot com>
- To: sjmunroe at vnet dot ibm dot com
- Cc: libc-alpha at sources dot redhat dot com
- Date: Wed, 28 Sep 2005 22:00:12 +0200
- Subject: Re: Make check failure with gcc-4.1 on PPC
- References: <433AF74D.9040901@us.ibm.com>
- Reply-to: Jakub Jelinek <jakub at redhat dot com>
On Wed, Sep 28, 2005 at 03:04:29PM -0500, Steve Munroe wrote:
> It seems that gcc-4.1 is getting getting more agressive using the
> powerpc fused-multiply-add (fmadd) which impacts the code gen (and
> rounding) for the bessel functions y1() and yn(). This causes
> math/test-float and math/test-double to report failures. This failure
> occurs for both powerpc32 and powerpc64.
>
> I reported this in bugzilla # 1384
> <http://sources.redhat.com/bugzilla/show_bug.cgi?id=1384> and provided a
> patch to libm-test-ulps as an attachment. The patch allows make check to
> pass with either gcc-3.4 or gcc-4.1.
IMHO patches should be mailed here. If you want to stick them to bugzilla too,
feel free, or better just post here and stick URL to the patch in
ml archives into bugzilla.
Jakub